Discussion about this post

User's avatar
Michael DAmbrosio's avatar

Kristen,

The lack of trust for many of us was Public Health being wrong in real time and not acknowledging or apologizing for the mistakes. Instead, they acted like Shamans and Astrologists, always defending their failed predictions with after-the-fact explanations and promoting unfalsifiable terror scenarios (YLE for example in fall 2021 made a widely circulated post on Facebook stating that PICUs would be overrun by 30% with kids which never happened, but parents pleaded to have schools canceled based on fear like this).

I’m sure you are a very nice person, mean well, but you too have engaged in this and I am sure you aren’t even aware. (making claims of being close to Herd Immunity, supporting mandates on a non sterilizing vaccines, promoting the vaccine in children while the rest of the world paused vaccinations on people under 55-65, etc).

I brought myself up to speed on your writings over the last 4 years on Covid going through your blog over the weekend (I have a lot of free time 😊). A lot of your early work resonated with me.

For example from your 5/10/20 piece “Lessons from Graduate School for the COVID Pandemic” you note:

“It is easy to only pay attention to the data that supports your hypothesis (and ignore

the data that goes against your hypothesis)” and “It is easy to get fooled by early data.”

On 2/9/201, you have an excellent post on “When you can never be wrong: the unfalsifiable hypothesis” outlining the perils of having a theory where counterfactuals can’t be constructed to test and refute a theory.

Great stuff, and I agree with all of that of course, as you are describing the tenets of scientific pursuit outlined by Merton, Feynman, Sagan, Randi, Shermer, etc.

Yet as I continued to go through your blog, I felt you were inadvertently committing the very mistakes you warn your readers against.

Consider on 10/9/21 your piece “If you’re vaccinated, why does it matter if I am not?” where you make an attempt to justify the vaccine mandates and make what I consider to be an unfalsifiable claim:

“I just spent a month in New England (which has a higher vaccination rate overall), and my risk of getting COVID there was much lower than my risk of getting COVID back home in Texas (which has a lower vaccination rate).”

Consider that on 10/9, the 7 day average for cases in Texas was 7232, against a population of 29 million. That means .025% is positive for Covid. Vermont, which was touted as the state where “no one was left to vaccinate” (and I choose Vermont too because Katelyn Jetelina referenced them in a post around 2021 as well I encouraged her to double check her claim [1]) – on October 9th, Vermont had 197 cases/day against 647K population, rate of .03%. Slightly more than Texas.

Why in that post did you believe that Texas was doing poorly while you were safe in New England? Doesn’t the most vaccinated state in New England having more cases per population than Texas falsify your claim? If it doesn’t, can your hypothesis even be falsified?

Further, consider that Covid would absolutely explode across the country and the world, following the vaccine rollout. You make claim in 16 different parts of your essay how “being vaccinated significantly reduces transmission”, yet Covid cases would increase 10-fold in both Texas AND Vermont within 3 months. The NYT graph for vermont: https://imgur.com/a/tptVuaR

Vermont would hit a peak of 2K cases/day, representing .31% of the population, while Texas would have an almost identical peak of 68K cases/day, at .23% of the population.

To me, that should have been enough to reject your claim that Texas was doing worse than New England and caused you to reconsider your entire hypothesis. Your appeals to base rate fallacy crumble when Covid is now up a magnitude post vaccine.

Especially since any confounder you search for to prop up this now tenuous claim favors Texas.

Population denominator Texas grew due to mass migration into the state. The baseline health of Texans is far worse than Vermont..ians? There was a surging undocumented migrant population pouring into Texas, further increasing the population denominator. Texans had tossed aside the masks and social distancing long before Vermont would… at this point you have to get creative to data dredge/p-hack your way into supporting your claim.

You may say “Fine, Covid is surging equally in Texas and Vermont, despite my claim Texas is worse off than New England, but it still greatly reduces your changes of severe outcomes”

That is also wrong.

Following your 10/9/21 post, Texas would have 22.6% more deaths than expected through the end of 2021, near identical than Vermont’s 23.1% - and again, the confounders you may search for all favor Texas having a better outcome.

I am not trying to get you in a “gotcha”. Brilliant people who know how science works make these mistakes ALL OF THE TIME. It’s easy to get fooled as Feynman notes in his takedown of Social Science as “A science which isn’t a science”.[2]

William Farr, a brilliant man and one of the greatest statisticians of the modern era, had all the data he needed to see that Snow’s hypothesis disproved Miasma theory for Cholera, yet Farr stuck to his favored hypothesis for 9 more years.

I could keep going through your blog, finding examples of you making claims that were easily falsified [3], but then I would succumb to Brandolini’s Law.

On a closing note, as I read your blog I see you inadvertently and repeatedly made the mistake of “one directional skepticism”. That is, the things you think are true get a free pass-regardless of how poor the supporting evidence – while only the ideas you disagree with do you apply skepticism.

Ivermectin and HCQ didn’t work, and you were right to explain why in your posts. Yet as noted in footnote [3], you fail to apply that same level of skepticism to the easily debunked claims in the Financial Times articles you cite that high vaccination rates are suppressing the spread of Covid.

A (to me) damning example of one-directional skepticism is your 7/22/20 post “Masked Science: Fact-checking Mask Studies” where you apply rigorous skepticism to a dozen studies claiming cloth masks don’t work, pointing out all the flaws in methodology, rigor, design, and conclusions.

Of course, the consensus on cloth and surgical masks has pivoted back to what we knew 1920-2019 – they don’t work – but you felt inclined to only debunk all the studies suggesting masks didn’t work. Why didn’t you eventually write a blog post admitting “hey we got cloth masks wrong – here is why”?

I wonder, why didn’t you apply that skepticism equally to the hundreds of nonsense studies claiming cloth masks were incredibly effective? Why not tear apart the 2 Stylists in a Missouri salon paper the CDC promulgated? The ridiculous Kansas Mask Study? Abaluck’s RCT in Bangladesh?

Katelyn has made the same mistake – any terrible study showing the wonders of cloth masks got a pass and added to her “Think masks don’t work? Look at the evidence” social media threads [4], while she suddenly became concerned with study design, confounders and p-values when an RCT suggested what we would find out to be true – they are largely worthless after all [5]

It seems to me, as a dispassionate scientist, that you entered the Social Media Covid Blog World with your mind made up – masks work, Covid vaccines work, lockdowns work, etc, and wound up stuck defending these extraordinary and failing theories for reasons I can only guess (Politics? Sunk Cost Fallacy? Availability bias?) using the very pseudoscientific practices you claim to reject.

I apologize if this comes off harsh, I am not trying to be mean spirited – there are just only so many characters we can put in a comment and I don’t know if any of this will be read anyway 😊

(PS 100% agree with your post “don’t censor threads” even if threads is dead – the message is spot on)

_______________________

[1] https://yourlocalepidemiologist.substack.com/p/state-of-affairs-europe-should-we/comments?s=r

[

2] Should be mandatory viewing for everyone, especially social scientists https://www.youtube.com/watch?v=tWr39Q9vBgo

“I might be quite wrong” – my mantra

[3] For example your 7/21/21 piece: “Some vaccinated people are getting COVID. What does this mean?” cites a FT article based on terrible data making a claim that the UK and Portugal are doing great by comparing them to sub-Saharan Africa – and of course right after this article UK and Portugal would explode in cases and deaths, while Africa would shrug and throw out unwanted vaccines.

https://x.com/nathanwpyle/status/1176860147223867393/photo/1

[4] It is fascinating to me to this day how after 100 years of not being able to find much benefit of masking against respiratory viruses, we conjured 35+ studies instantly showing how amazing they are in a few months

https://www.facebook.com/permalink.php?story_fbid=202002698114314&id=101805971467321

[5] https://www.facebook.com/permalink.php?story_fbid=pfbid01SraZhfgd3fvJ969SrsjufxCuWE92DJRTW8dXbYqMFemjX5WYVUwD85GPJbMFVQyl&id=101805971467321

Expand full comment
Greg Lanman's avatar

I think we made assumptions, that everyone would see us as SMEs and that they would "understand" and comply with our advice. No one expected that the POTUS would openly challenge our expertise. No one expected that several Red State Governors would openly challenge our expertise. I don't have a counterargument for politicians who have little or no respect for science. I don't have any counterarguments for politicians who do not have the health, safety, and welfare of their citizens as their primary motivation during a pandemic. I know that we did the best, we could do, given the circumstances. But nothing changes the fact that (maybe) 3M Americans died needlessly during the pandemic, mostly because they either choose to ignore us, because they believed politicians over the "expertise" or politicians muddy the water enough, to confuse a significant number of Americans; that they decided that no action was their best option.

Expand full comment
79 more comments...

No posts